The freeze-fractured membranes of subpial (marginal) and perivascular astrocytes contain orthogonal arrays or assemblies of small particles as do ependymal cells. Assemblies do not occur in choroid plexus epithelium. The functions of the assemblies are unknown but might have something to do with secretion or absorption of fluid and solutes. Solutions of 10 to the minus 4th power N ouabain and of acetazolamide, which reduces CSF production, are being infused ventriculo-cisternally in mice. We are also attempting to change the number or configuration of the assemblies by causing marginal astrocytes and ependyma to proliferate or hypertrophy after painting 30% to 66% gels of aluminum hydroxide onto the floor of the IV ventricle and the dorsal surface of the medulla. In order to determine whether such assemblies are present in astrocytes and ependyma cultured in vitro we must first identify the cells as glial or fibroblastic. To make this distinction, a cytoimmunochemical technique is used whereby fixed cells in situ and in vitro, are labeled with an antibody to a "non-neuronal protein" (NNE) which, we have found, labels astrocytes, ependyma and choroid epithelium but not neurons. BIBLIOGRAPHIC REFERENCE: Prescott, L. and Brightman, M.W.: The sarcoleman of Aplysia smooth muscle in freeze-fracture preparations. Tissue and Cell 8: 241-258, 1976.